Reconciling Observed Patterns of Temporal Occurrence with Cladistic Hypotheses of Phylogenetic Relationship

Cladistic analyses of well-sampled groups with a complete and abundant fossil record can yield phylogenetic hypotheses that conflict with stratigraphic data, even to the extent of supporting phylogenies that appear to invert the stratigraphy.This is most probably due to the convergent evolution of similar morphologies (7.e. homoplasy), rather than the inadequacy of the fossil record.Several ways have been proposed to resolve this problem (stratophenetics, stratocladistics, character refinement, etc.).This paper proposes an iteractive technique to construct a tree more consistent with the observed fossil record by building separate phylogenies for different stratigraphic intervals which can then be assembled into a composite phylogeny.Columbellid gastropods of the genus Strombina Morch, 1852, were used to test this approach.Strombina originated and diversified in the Caribbean during the Miocene and Pliocene.During the Pliocene, they became nearly extinct in the Caribbean, but diversified in the eastern Pacific.Phylogenies of 42 species based only on shell morphology (49 characters, 186 states) yielded trees with high stratigraphic inconsistency and ghost lineages that require the presence of descendants 10 million years or more before the first appearance of their hypothesized ancestors.Removal of species that originated after the Pliocene resolved most of these stratigraphic inconsistencies although some ghost lineages remained.This Miocene/Pliocene tree was then used to root the trees illustrating the relationships among Pleistocene/Holocene species.This final composite tree is more consistent with the hypothesized fossil record for the group than the original tree.